Concurrent Submit | Zhenzhen_Flux_Kontext_Edit

Where the other Kontext nodes generate or edit from a single reference, this one is the dedicated edit entry point: source image in, instruction in, edited image out - and it lets you produce 1 to 4 variants per call, which is exactly the shape of "I need to see this in four styles before I pick one." The Comfyui-zhenzhen pack resells BFL's Kontext API (dev/pro/max tiers) with its own key, and the Concurrent Submit wrapper means you can stack a batch of these edits rather than running them one at a time.

Kontext's trick, from the KB panel: it consumes reference images in-context instead of through adapter embeddings, so identity holds across successive edits. That's why it's the classic turnaround-sheet and outfit-variation tool. This node hands that behavior to your ComfyUI graph over HTTP.

How the concurrent bit works

The plain Comfly_Flux_Kontext_Edit node blocks until the edit lands. The Submit variant runs it in the pack's shared image pool (30 workers) and returns a task handle immediately. Collect via Concurrent Collect Images (30): wire task_1 … and the collector waits for the whole batch, returning images in slot order plus a status string. failure_mode = fail-fast (default) or placeholder.

Batch math: eight prompts × four variants = thirty-two potential edits. In one run, in parallel. That's the pitch.

The inputs that matter

• prompt (required) - the edit instruction.
• image - your source image to edit.
• model - flux-kontext-dev, -pro, or -max. Dev is the cheapest of the three and often plenty for edits; max is the top tier. (Per the README, dev was added to this edit node in the same June update that brought it to the base Kontext node.)
• aspect_ratio - 21:9 → 9:21, default 1:1.
• num_of_images - 1–4 variants per call. Remember the API meters per image, so four variants is four charges.

Output: one task (COMFLY_IMAGE_FUTURE) → Collector → the edited image(s).

Installing it

Ships in Comfyui-zhenzhen:

• ComfyUI Manager → Install Custom Nodes → search "Comfyui-zhenzhen" → install → restart.
• Or cd ComfyUI/custom_nodes && git clone https://github.com/T8mars/Comfyui-zhenzhen and restart.

No model files, no VRAM cost. You need a paid token from the author's storefront (ai.t8star.org), pasted into api_key.

Gotchas

The num_of_images > 1 multiplies your spend - cheap for dev, less cheap for max, so don't leave it at 4 out of habit. The README notes this edit variant doesn't charge a separate image-upload fee the way some upload-based nodes do (a small thing, but nice when you're iterating). Standard pack warnings still apply: the overseas endpoint needs a proxy from mainland China, and a 500 usually means upstream load - retry before you tear the graph apart.